The Great IT Triage: How AI’s Rising Price Tag Is Reshaping Next Year’s Budgets
Picture a CIO in October staring at a spreadsheet where the AI line item has swelled past every forecast from a year ago. Everything else on the page — the server refresh, the help desk headcount, the “nice to have” analytics tool — is frozen, trimmed or gone. This is the defining budget exercise of the year for technology leaders across industries: not “How do we add AI” but “What do we cut to afford it.”
The numbers explain why. Global IT spending is forecast to reach $6.31 trillion in 2026, up 13.5% year over year — the highest growth rate in a decade — driven primarily by AI infrastructure investment. AI-specific spending alone is growing 47% year over year to $2.59 trillion as organizations move from pilot projects to production deployments. That growth is not evenly distributed: Data center systems are climbing 55.8%, while software spending is up 15.1% and devices just 8.2%. AI is actively reshaping the entire hierarchy of technology spending.
Why costs have gotten so steep
Three main forces are compounding. First, compute: training and running AI models at scale requires GPU capacity that remains scarce and expensive, and cloud inference bills climb with every user interaction. Second, licensing sprawl: AI features are being bolted onto existing software contracts including Copilot, Einstein and dozens of similar programs, often at a per-seat premium that stacks on top of what companies already pay. Third, talent: AI-skilled engineers now command skill premiums of 20-40% over standard technical roles, and that scarcity pricing shows no sign of easing.
However, there is a fourth, less visible cost too. Rising security spending is directly tied to the expanded attack surface that generative AI introduces. AI does not just cost money to run. AI creates downstream obligations in governance, data cleanup and risk management that rarely show up in the original pitch.
What’s getting triaged
The result is a budget reallocation happening inside IT departments everywhere, even ones with growing overall spend. 43% of CIOs report their operating budgets increased this year and another 38% say budgets held steady, so this is not primarily a story of shrinking IT budgets. This is a story of what growth is being spent on, and what is being sacrificed to fund it. AI is becoming a material portion of IT budgets, often targeted at 15-20% of the total, a share that must come from somewhere.
There is a triage pattern showing up in real time: Modernization and security spending isn’t disappearing, but it’s increasingly framed as a cost of utilizing AI, not a separate priority competing for its own dollars.
Elsewhere, the tradeoffs are blunter. Technical debt remediation is now cited by 34% of CIOs as the top barrier to AI adoption, meaning legacy modernization work that used to stand on its own is being reframed and often delayed as a prerequisite to something else, rather than funded for its own sake. Meanwhile, FinOps and cost optimization have climbed the priority list specifically because cloud and AI spending are growing together, forcing IT leaders to build new cost-monitoring functions just to keep the AI bill from spiraling further.
The scrutiny cuts both ways
Even AI spending itself isn’t immune from the belt-tightening mindset. New, non-AI initiatives are facing tighter review, and technology leaders are being asked to show faster returns and lower risk. AI budgets may be growing, but they are no longer insulated from scrutiny. The era of funding AI experimentation on faith is ending. Every dollar, AI or not, increasingly has to justify itself against a shorter payback window.
Some IT leaders argue this discipline is overdue, that AI’s cost has forced a rigor onto technology spending that was missing for years, killing pet projects that never had clear ROI. Others warn that squeezing foundational work like security and infrastructure to chase AI’s promise is a bet that could backfire, leaving companies with flashy AI features sitting on brittle, undersecured systems. Cybersecurity spending is rising too; technology leaders are increasing security budgets above inflation, but whether that keeps pace with AI-driven risk is an open question, not a settled one.
What smart triage looks like
For companies navigating next year’s budget cycle, a few principles are coming out as sound practice regardless of where you land on that debate. Tie AI spending to measurable outcomes before scaling further, rather than funding based on momentum alone. Treat core security and data infrastructure as non-negotiable instead of a flexible line item that absorbs AI’s overflow. Avoid multi-year vendor lock-in while AI pricing remains this volatile — the deal that looks good today may look expensive in eighteen months. And build governance that spans departments, so that AI spending doesn’t quietly become shadow IT scattered across a dozen expense reports.
None of this means AI investment is a mistake. The productivity case is real, and the CIOs pouring budget into it aren’t wrong to do so. But triage, by definition, means someone doesn’t get treated first. The question every technology leader needs to answer honestly this budgeting season is what they are comfortable leaving behind to invest in AI, and whether they have actually reckoned with the cost of that choice.
Jason Rosenfeld
Jason Rosenfeld is a recognized ServiceNow executive veteran with over 15 years of experience driving innovation and success within the ServiceNow community, and 25 years of experience leading professional services and consulting organizations. As the Chief Growth & Alliances Officer at NewRocket, Jason is responsible for orchestrating the company’s global go-to-market strategy, cultivating high-impact relationships with ServiceNow, and accelerating revenue growth through strategic account planning and innovative partnerships.